Based on my project, I am getting started to review the ASR course from CMU.  This will help me understand what happens under the hood. To be honest, the slides from CMU is among the best that I have ever seen for ASR courses. This post will do a short review of the Lec 1…

CNTK Official Installation Error

CNTK is a really good AI framework that maintained by Microsoft. Its competitors are Tensorflow,  Torch and other popular ones used both in academy and industry. Well, big company right? But I have installed CNTK almost ten times. And I found that there are two errors in the official doc that is always there. I…

MSc Final Dissertation Note (Kaldi Installation)

I am doing my final project that doing experiments about the function of Dropout in the LSTM for TIMIT corpus. Kaldi and CNTK are the tools chosen for this one. Kaldi is a really powerful tool for decoding waveform files into features. The prerequisites for this are to install Kaldi and its relative dependencies the…

TIMIT Introduction (the Official Doc)

The DARPA TIMIT Acoustic-Phonetic Continuous Speech Corpus (TIMIT) Training and Test Data NIST Speech Disc CD1-1.1 The TIMIT corpus of read speech has been designed to provide speech data for the acquisition of acoustic-phonetic knowledge and for the development and evaluation of automatic speech recognition systems. TIMIT has resulted from the joint efforts of several…

Leetcode 357. Count Numbers with Unique Digits

This post I will make a short review about Python data structure by solving the problem of Leetcode 357. The problem is as below: Given a non-negative integer n, count all numbers with unique digits, x, where 0 ≤ x < 10n. Example: Given n = 2, return 91. (The answer should be the total…

Extreme Computing (Hadoop Map Reduce) 4

This one will continue the discussion for the second assignment of Extreme Computing in the University of Edinburgh Task 4 For this task you will use a dataset from StackOverflow and extract specific pieces of information. Initially, you should understand the format of the dataset, next you will need to do parse each post, and…

Extreme Computing (Hadoop Map Reduce) 3

This blog will be the second assignment of Extreme Computing from the University of Edinburgh. Extreme Computing Second assignment Based on an Assignment by Michail Basios and Stratis Viglas In this assignment, you will address real-world cases where MapReduce can be used. Initially, you will deal with a problem related to information retrieval: you will…

Extreme Computing (Hadoop Map Reduce) 2

In this part, we will still discuss the remaining problems solved in the coursework of Extreme Computing in the University of Edinburgh.  If you would like to see the whole introduction, please go to Part 1. Task 5 Create a version of the two-word counting program that uses a combiner. Is it faster? Mapper (Python)…

Extreme Computing (Hadoop Map Reduce) 1

Now our network generates more and more data as time goes by. Computer scientists are trying to understand the pattern under all of the raw data. Hadoop could be used to clean data and send then results to the pipeline, or do some SQL style manipulations or many more advanced tasks on different types of data. Today…

Hello World

Hi there, I am a student majoring Machine Learning, AI in the School of Informatics at the University of Edinburgh. The city is beautiful and people are really nice here. I’d like to share you some ideas of machine learning, deep learning,  computer vision and robot motor control. Here are some awesome websites bout machine…